Boost logo

Boost :

From: Beman Dawes (bdawes_at_[hidden])
Date: 2002-08-06 08:15:51


The new regression test reporting [optionally] makes compiler warnings
visible. That seemed to be a Good Thing, but it isn't 100% obvious which
warnings we should be chasing.

It seems to me a policy might go something like this:

There is an agreed upon set of warnings that:

    -- we will enable in as many compilers as possible.
    -- we will try to clear when they appear in tests.
    -- an individual test may disable.

There is an agreed upon set of warnings that:

    -- we will disable in as many compilers as possible.
    -- we will ignore when they appear in tests.
    -- an individual test may enable.

At first, we might start out with a very small list of desired warnings,
and then expand it as we gain experience. This to avoid initially being so
overwhelmed with warnings that we just ignore them.

Comments? Suggested initial list of enabled warnings?

--Beman


Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk